sign in sign up
<<
Home , Speleothem

International Journal of 48 (1) 63-74 Tampa, FL (USA) January 2019

Available online at scholarcommons.usf.edu/ijs International Journal of Speleology Off icial Journal of Union Internationale de Spéléologie

Cave dripwater isotopic signals related to the altitudinal gradient of Mount-Lebanon: implication for speleothem studies Carole Nehme1,2,3*, Sophie Verheyden2,4, Fadi H. Nader5, Jocelyne Adjizian-Gerard6, Dominique Genty7, Kevin De Bondt2, Benedicte Minster7, Ghada Salem3, David Verstraeten2, and Philippe Claeys2 1Laboratoire IDEES UMR 6266 CNRS, University of Rouen Normandy, rue Thomas becket, 76821, Mont-Saint Aignan Cedex, France 2Analytical, Environmental & Geo-Chemistry, Department of Chemistry, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Elsene, Brussels, 3ALES, Association Libanaise d'Etudes Speleologiques, Mansourieh El-Matn, Lebanon 4Directorate of Earth and Quaternary, Royal Belgian Institute of Natural Sciences (RBINS), rue Vautier 29, 1000, Brussels, Belgium 5Energie France Pétrole-Nouvelles, 4, avenue de Bois-Préau, 92852, Rueil-Malmaison Cedex, France 6CREEMO, Saint-Joseph University of Beirut, Faculty of Human Sciences, Rue de Damas, BP 17-5208, Beirut, Lebanon 7Laboratoire des Sciences du Climat et de l’Environnement (LSCE/IPSL), UMR 8212 CEA/CNRS/UVSQ, F-91191, Gif sur Yvette Cedex, France

Abstract: An important step in paleoclimate reconstructions based on vadose carbonate deposits or speleothems is to evaluate the sensitivity of the cave environment and speleothems to regional climate. Accordingly, we studied four , located at different altitudes along the western flank of Mount-Lebanon (Eastern Mediterranean). The objectives of this study are

to identify the present-day variability in temperature, pCO2, and water isotopic composition and to assess the possible influence of the altitudinal gradient on cave drip waters and cave streams. We present here an overview of the spatial variability of rainwater based on local

and regional data, and we compare these data with our results, i.e., temperature, air pCO2, and the isotopic composition of cave water and modern cave collected in 2011 and 2014. The results show that the rainwater isotopic signal is generally preserved in the cave dripwater isotopic composition with some exceptions in large caves with high ceilings where evaporation effects may influence its isotopic composition. The altitude effect observed in rainwater isotopic composition seems to be transferred to the cave dripwater. Different δ18O/100 m gradients between dripwater and rainwater (0.13‰ and 0.21‰, respectively) are noted. This is mainly attributed to the δ18O/100 m value of the dripwater which is site-specific and dependent on i) local processes within the epikarst/soil, ii) the relation to the precipitation altitude gradient and iii) the extension of the defined infiltration basin.

Keywords: drip water, isotopic signal, Lebanon, caves, altitude gradient Received 8 February 2019; Revised 26 February 2019; Accepted 26 February 2019

Citation: Nehme C., Verheyden S., Nader F.B., Adjizian-Gerard J., Genty D., De Bondt K., Minster B., Salem G., Verstraeten D. and Claeys P., 2019. Cave dripwater isotopic signals related to the altitudinal gradient of Mount-Lebanon: implication for speleothem studies. International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) ISSN 0392-6672 https://doi.org/10.5038/1827-806X.48.1.2253

INTRODUCTION Rainwater and dissolved carbon circulate through the unsaturated zone, i.e. the upper part of the epikarst, Speleothems, which are secondary cave carbonate which is affected by dissolution and is characterized deposits that precipitate from cave drip water are by a mainly vertical transfer of percolation water to increasingly used to reconstruct changes in regional the cave (Hendy, 1971; Bar-Matthews et al., 1996; climate and vegetation. Their isotopic composition, Ford & Williams, 2007; Fairchild & Baker, 2012). δ18O, and δ13C is influenced mainly by respectively the Several cave monitoring programs have been isotope signature of rainwater linked to temperature conducted worldwide, providing information on the (Clarck & Fritz, 1997; Lachniet, 2009) and by the role of local cave environment and hydrology that carbon isotopic composition of dissolved carbon possibly influence -based palaeoclimate influenced by the soil bioactivity (δ13C), linked to records (Bar-Matthews et al., 1996; Spotl et al., vegetation and thus to temperature and water 2005; Baldini et al., 2006; Verheyden et al., 2008a; availability (Hellstrom et al., 1998; Genty et al., 2006). Mattey et al., 2010; Miorandi et al., 2010; Tremaine

*[email protected] The author’s rights are protected under a Creative Commons Attribution- NonCommercial 4.0 International (CC BY-NC 4.0) license. 64 Nehme et al. et al., 2011; Johnston et al., 2013; Genty et al., 2014; THE STUDY AREA: REGIONAL CLIMATIC Deininger et al., 2014; Van Rampelbergh et al., 2014; CONTEXT AND SITE DESCRIPTION Suric et al., 2016; Beddows et al., 2016). These studies aim at understanding better the registration of the The Levant region in general is mainly influenced nowadays climatic signal in speleothems in order to by the mid-latitude westerlies (Fig. 1A), which more precisely constrain the past climatic records. originate from the Atlantic Ocean, forming a series of In the Levant region, although speleothems subsynoptic low-pressure systems (Gat et al., 2003; from Palestine/Israel are well studied and proxy Ziv et al., 2010) across the Mediterranean Sea. In interpretations well-supported by monitoring winter, cold air plunging south over the relatively programs, the rest of the region is still understudied warm Mediterranean enhance cyclogenesis, creating in terms of speleothem-based paleoclimate studies the Cyprus Low (Alpert et al., 2005). This low- (Nader et al., 2007, Verheyden et al., 2008b, Cheng pressure system drives moist air onshore, generating et al., 2015, Nehme et al., 2015; 2018). Nearly no intense orographic rainfall across the mountains of cave monitoring information is available and only the northern Levant. The duration, intensity, and limited rainwater isotopic data is available (Aouad- track of these storm systems strongly influence Rizk et al., 2005; Gat el al., 2005, Abou Zakhem & the rainfall amount in this region. In summer, the Hafez, 2010). Aouad-Rizk et al. (2005) and Koeniger westerly belt is shifted to the north, following the & Margane (2014) each defined a local meteoric northern shift of the North-African subtropical high water line (LMWL) based on data from central Mount- pressures, and the region experiences hot and dry Lebanon. Both studies display some differences which conditions with more southward winds. In Lebanon will be further discussed in the paper. The Levant (Fig. 1B and 1C), the annual rainfall varies between (East-Mediterranean) is characterized by abrupt 700 and 1000 mm along the coastline and more than temperature and rainfall gradients, due to its current 1400 mm in higher mountains with 4 months snow location on the arid/semi-arid boundary and to its coverage (Shabaan et al., 2015). As a consequence of steep topography between coastal and inland areas. the above circulation system, the climate is seasonal In a first attempt to understand the local with wet winters (November to February) and dry, hot environmental conditions, four Lebanese caves summers (May to October). A general N-S gradient in were investigated for their temperature, pCO2 rainfall amount and mirrored by the isotopic signal concentration, and dripwater isotopic composition. (δ18O and δ2H) is clearly evident from northern Syria Here, modern changes in the rainwater isotopic (Abou Zakhem & Hafez, 2010), to southern Israel/ composition across the steep altitudinal trend of Palestine (Gat et al., 2005). A West-East gradient, Mount-Lebanon, are compiled based on a literature i.e. from the Levantine coastline to inner regions review, and compared with the observed changes in (Fig. 1A), is also visible as a consequence of the modern cave drip waters. The main objectives of this continental and/or altitudinal effects related to the paper are to discuss if rainwater signal is generally Rayleigh distillation processes (Dansgaard, 1964; preserved in the cave dripwaters and to assess the Rozanski et al., 1993). possible influence of the altitudinal gradient on cave Four caves are selected at different altitudes along drip waters and cave streams. This study will help a transect from the coast to the Makmel Mountain, also verify the ability of cave waters in Lebanon to which is the highest peak in the Mount-Lebanon transfer spatial changes in isotopic composition of range (Fig. 2). These are: Kanaan Cave (96 m above rainwaters and by extension of spatial and temporal sea level - asl), Jeita Cave (98 m asl), Mabaage changes in regional climate. Cave (770 m asl), and Qadisha Cave (1720 m asl).

Fig. 1. Climate and geographic setting of the study area. A) Eastern Mediterranean map showing the position of the mid latitude winds (http://iridl. ldeo.columbia.edu/Maproom), NS and EW precipitation gradients and δ18O mean values NS and EW precipitation gradients, of rainwater stations over coastal and inner cities (Kailani et al., 2003; El-Asrag, 2004; Aouad-Rizk et al., 2005; Dirican et al., 2005; Gat el al, 2005; Saad et al., 2005; Abou Zakhem & Hafez, 2010; GNIP database); B) Precipitation gradients of Lebanon and histograms of Beirut and the Cedars with mean annual rainfall and temperature (Abi-Saleh & Safi, 1988; http://fr.climate-data.org); C) Rainwater isotope graph with several published meteoric waterlines: the Lebanese MWL in Saad et al. (2005), the Global MWL in Rozanski et al. (1993), and the Mediterranean MWL in Gat (1980).

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 Caves dripwater isotopic signals at Mount-Lebanon: implication for speleothem studies 65 Except for Qadisha Cave which is developed mainly dry and active galleries (Karkabi, 1990), a permanent in Quaternary deposits and Cretaceous stream with a discharge of 1 to 25 m3/s (Doummar, (Dubertret, 1975), Mabaage, Jeita and Kanaan caves 2012) and is the most visited in Lebanon. develop in the middle Jurassic Kesrouane Formation, A 75 m deep canyon connects fossil galleries with the a faulted micritic and sequence lower galleries in the downstream extremity of the (Walley, 1998). The studied caves (Table 1) are located 10 km karstic network, making the cave a well- in the western flank of Mount-Lebanon (Fig. 1) ventilated system (Fig. 2). Mabaage Cave 400 m along a N-S altitudinal transect. All four caves were long, located at 40 km northeastern of Beirut and in previously studied for their speleothem content the inner part to the Fidar valley (Jabbour-Gedeon (Verheyden et al., 2008a; Cheng et al., 2015; Nehme & Zaatar, 2013) was recently transformed into a et al., 2015, 2018). touristic cave during summer but closes in winter Kanaan Cave (162 m long) is located 15 km due to flooding of the cave stream. Finally, Qadisha northeast of Beirut. This fossil cave was discovered Cave, located in the northern part of Mount-Lebanon, after quarrying activity in the late 1990s (Nehme et hosts a permanent with a discharge rate up al., 2009). The Jeita multi-level system cave, located to 1 m3/s (Edgell, 1997). Qadisha cave was partially at 4.5 km distance from the coast, hosts a series of transformed into a touristic cave in 1934.

Fig. 2. The study area and cross-sections of A) Kanaan Cave (Nehme et al., 2013); B) Jeita Cave (Karkabi, 1990; Nehme, 2013); C) Mabaage Cave (Zaatar et al., 2013). For the cross-sec`tion of Qadisha Cave, see Tawk et al. (2008). Red stars indicate the location of water samples taken inside each cave and cave elevations are in m above sea level (asl). Refer to Table 1 for additional cave site data. The vegetation cover above the caves mainly trees) growing on calcareous slopes above Kanaan, develops in shallow Mediterranean soil. Between 100 Jeita and Mabaage caves. The vegetation cover above and 800 m asl, the vegetation consists of densely Qadisha Cave (1720 m asl) is composed of sparse evergreen shrubs (juniper, oaks, and partially pine herbs, shrubs, and conifers (Table 1).

Table 1. Locations, morphology of the studied caves, and their soil characteristics. Infiltration Entrance Length Cave Coordinates Cave type basin Aspect Host Rock Vegetation type (m) (m) elevation (m) Horizontal, dense garrigue, Kanaan 33°54'25"N; 35°36'25"E 98 547 SE-NW 162 J4-J5 relict pine forest

Horizontal, dense garrigue, Jeita upper 33°56'35"N; 35°38'48"E 96 1067 N-S 1300 J4-J5 relict pine forest

Horizontal, dense garrigue, Jeita Lower 33°56'35"N; 35°38'48"E 60 1669 E-W 8750 J4-J6 active pine forest

Descending sparse garrigue, Mabaage 34°06'25"N; 35°46'01"E 770 1379 E-W 400 J6 cave oaks, pine

horizontal, Sparse hurbs, Qadisha 34°14'38"N; 36°02'11"E 1720 2244 NE-SW 1076 Q; C4 spring shrubs, conifer

SAMPLES AND METHODS September and November 2014 in Jeita, Qadisha, Mabaage, and Kanaan caves.

A total of 35 cave drip water and 12 underground Temperature and pCO2 of cave air were measured stream water samples were collected in the Jeita and using a hand thermometer with a precision of Qadisha caves for δ18O and δ2Η analyses, respectively. 0.5°C and a Dräger pump system (σ ± 50 ppmv), The samples were obtained during two sampling respectively. Continuous temperature monitoring campaigns: a first one held in September 2011 in using a Niphargus (Burlet et al., 2015) temperature Jeita and Qadisha caves and a second one between logger (precision of 0.1°C and resolution of 0.05°C)

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 66 Nehme et al. was pursued from December 2015 to March 2017 in surface watershed. The underground waterflow main Qadisha and Jeita caves with one measurement every directions identified in Hakim (1985) and Hakim et al. 20 minutes. (1988) for the Lebanese basins were considered Isotopic analyses of cave waters collected in to derive the most significant infiltration surface 2014 were carried out using a PICARRO L2130-i above the caves. The mean altitude is then calculated Cavity Ring-Down Spectrometer (CRDS) at the Vrije for the delimited infiltration basin for each cave using Universiteit Brussel. Measured values were corrected the DEM. Note that the Jeita Cave develops on two- using three house standards calibrated against the levels (an upper fossil and a lower active gallery) thus international VSMOW2, GISP, and SLAP2 standards has two different infiltration elevations. following the method described in De Bondt et al., (2018). Analytical uncertainties (2σ) equal 0.06‰ for RESULTS δ18O values and 0.3‰ for δ2H values. Water samples collected were analyzed in 2011 at the Laboratoire Cave air temperature and pCO2 des Sciences du Climat et de l’Environnement (LSCE- Cave air and underground stream temperatures, CEA), Paris. Hydrogen isotopes were measured on an measured at different sites inside each cave (see ISO-PRIME mass spectrometer and a PICARRO CRDS Supplementary Data), a fairly constant (Fig. 3A) with with a 1 sigma error of ±0.7‰. Oxygen isotopes were variations of less than 1°C over the sampling period. analyzed using a Finnigan MAT 252 by equilibration The measured air temperatures in Kanaan (19°C ± 18 with CO2. The 2 sigma error of the δ O is ±0.05‰. All 0.5), Jeita upper (20°C ± 0.5), Mabaage (13°C ± 0.5), values obtained from both laboratories are calibrated and Qadisha caves (9°C ± 0.5) all display autumn against and reported in permill (‰) relative to Vienna values roughly in agreement with the outside mean Standard Mean Ocean Water (V-SMOW2). temperature (Fig. 3B) data (Karam, 2002), despite To calculate the altitudinal gradient of the cave the small offset compared to the surface temperature dripwaters with respect to the altitude of the entrance trendline and some small internal changes (up to and the infiltration basin of the studied caves, the 0.3°C) as shown by the continuous monitoring data infiltration basin elevation (Table 1) was derived in Jeita and Qadisha (Fig. 3C). after plotting the georeferenced caves maps on a As for the pCO2 concentrations in each cave, the Digital Elevation Model (DEM) using a Geographical measured values reached 3,600 and 8,000 ppmv in Jeita Information System (ArcGIS). The infiltration watershed and Mabaage caves respectively, whereas low values area of the cave is defined by considering the altitudes (600 ppmv) close to the atmospheric concentrations are between the cave entrance and the limit of the detected in Qadisha Cave (see Supplementary Data).

Fig. 3. A) Cave-air and underground stream temperature measurements collected each month in both 2011 and 2014 171 campaigns; B) cave- air temperature trendline (black line) vs the cave entrance altitude. The air-temperature trendline from compiled data of Lebanese meteorological stations (Karam, 2002) is represented in red line; C) continuous cave-air temperature monitoring from November 2014 to March 2017 for Qadisha and from January 2016 to March 2017 for Jeita Cave. Cave dripwater δ18O and δ2H Qadisha Cave, located at the highest altitude in our Cave dripwaters and stream waters δ18O and study area, shows an average of -8.48 ± 0.05‰ for δ2H are summarized in Table 2 and detailed in the δ18O and –46.1 ± 0.31‰ for δ2H with an amplitude Supplementary Data. Jeita Cave dripwaters exhibit of 0.2 and 1.1‰, respectively. The variability of the an average of -5.7 ± 1.1‰ for δ18O and –26.6 ± 6.9‰ dripwater oxygen isotopic signal in Kanaan (avg. for δ2H (Table 2), with an amplitude of 2.9 and 20.5‰, -5.4‰) and Qadisha (avg. -8.5‰) does not exceed respectively. As for Kanaan Cave, measurements show ±0.1‰ (Table 2). Drip water values in both Jeita an average of -5.40 ± 0.04‰ for δ18O and –24.0 ± 0.2‰ and Qadisha caves show lower isotopic values than for δ2H (Table 2), with an amplitude of 0.2 and 0.5‰ Jeita (avg. -7.2‰) and Qadisha (avg. -9.0‰) stream respectively. Mabaage Cave located at higher altitude waters. However, the difference is much higher (770 m) shows an average of -7.2 ± 0.6‰ for δ18O and between dripwater and stream water isotopic values –36.6 ± 6.2‰ for δ2H, whereas the amplitude varies in Jeita Cave (~1.5‰) than those of Qadisha Cave between 1.8 and 14.0‰, respectively. (~0.4‰).

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 Caves dripwater isotopic signals at Mount-Lebanon: implication for speleothem studies 67 Table 2. Summary of the isotopic results of the drip and stream waters collected from the studied caves. Note that Jeita upper is the fossil cave and Jeita lower in the active cave with a permanent stream. (n) is the number of samples and oxygen isotopic composition ranges. The mean (avg.), maximum and minimum values of the δ18O drip and stream water and mean (avg.) of the δ2H of the autumn values, are reported in ‰ VSMOW.

Fossil Caves n Dripwater δ18O (‰ VSMOW) Dripwater δ2H (‰ VSMOW) min max avg. 2s avg. 2s Kanaan 6 -5.48 -5.37 -5.43 ±0.04 –24 ± 0.2 Jeita Upper 11 -6.92 -4.05 -5.71 ±1.11 –26.6 ± 6.9 Mabaage 8 -8.28 -6.64 -7.18 ±0.66 -36.6 ± 6.2 Qadisha 10 -8.55 -8.38 -8.48 ±0.05 -46.1 ± 0.3 Cave Streams n Stream water δ18O (‰ VSMOW) Stream water δ2H (‰ VSMOW) min max avg. 2s avg. 2s Jeita lower 6 -7.35 -7.17 -7.28 ±0.06 -35.73 ± 0.48 Qadisha 6 -8.96 -8.92 -8.95 ±0.02 -49.24 ± 0.26

DISCUSSION (Table 3). The trendline slopes of the MWL (Lebanon, Mount-Lebanon, etc.) are different than that of the In order to determine if the current spatial gradients Mediterranean MWL, due mainly to a secondary in rainwater isotopic composition are recorded in the evaporation effect during rainfall events (Saad et al., cave dripwater, we discuss i) the available meteoric 2005; Saad & Kazpard, 2007). The evaporation occurs water lines of Lebanon and their altitudinal trends, mostly during hot (dry) seasons and is particularly ii) the cave waters δ18O/δ2H signals compared to impacting light rains. Consequently, this process will the available δ18O/δ2H rainwater data, and iii) the determine the lowering of the slope and the “d-excess” altitudinal trend in rainwater δ18O/δ2H and the value of the rain sample (Clark & Fritz, 1997). potential altitudinal trends in cave water δ18O to test In general, the constructed Lebanese Meteoric water for their agreement. lines based on δ18O and δ2H data of rainwater are roughly in agreement with a general depletion trend Rainwater data of Lebanon: different meteoric with elevation. However, the local MWL (Koeniger & water lines and altitudinal trends Margane, 2014; Koeniger et al., 2017) for the Kelb Several studies on the rainwater isotopic signal basin, the Mount-Lebanon (Aouad-Risk et al., 2005) in Lebanon (Aouad-Rizk et al., 2005; Saad et al., and the general Lebanese MWL (Saad et al., 2005; Saad 2005; Saad and Kazpard, 2007; Koeniger & Margane, & Kazpard, 2007) are calculated based on different 2014; Koeniger et al., 2017) exist in the literature locations of the meteorological stations (Fig. 4A).

Table 3. Summary of the Global MWL and the MWLs previously calculated for the Mediterranean, Lebanon and Kelb basin. Meteoric water line Equation Author Global meteoric water line δ2H = 8* δ18O + 10 Craig, 1961 Mediterranean water line δ2H = 8* δ18O + 22 Gat, 1980; Gat et al., 2003 Lebanese water line δ2H = 7.13* δ18O + 15.98 Saad et al., 2005 ; Saad & Kazpard, 2007 Mount-Lebanon water line δ2H = 6.3* δ18O + 8.2 Aouad-Rizk et al., 2005 Kelb basin water line (Local) δ2H = 6.04* δ18O + 8.45 Koeniger & Margane, 2014; Koeniger et al., 2017

The MWL after Aouad-Risk et al. (2005) referred Cave waters δ18O/δ2H signals compared to the here as the Mount-Lebanon MWL, is constructed available δ18O/δ2H rainwater data using data from meteorological stations which display The isotopic results of cave waters (drip and stream) the same E-W trend than the stations used for the of the four studied caves fall well on the Mount MWL of the Kelb basin (Koeniger & Margane, 2014). Lebanon MWL (Aouad-Risk et al., 2005), except for Indeed, the MWL after Aouad-Risk et al. (2005) and some of the Jeita Cave dripwaters (Fig. 4C). In general, 18 the local MWL after Koeniger & Margane (2014) show Kanaan, Jeita, Mabaage and Qadisha δ Owater (drip the same gradient (Table 3). and stream) values seems to fall more closely to the The altitudinal trendline (Fig. 4B) used in this study Mount Lebanon and Lebanese MWL than the regional is constructed after the latest rainwater data (Koeniger Mediterranean MWL (Gat, 1980; Gat et al., 2003). 18 & Margane, 2014; Koeniger et al., 2017) from stations The δ Odrip values of Kanaan Cave are at the lower located in the Kelb basin (central Mount-Lebanon) part of both Lebanese and Mount-Lebanon MWLs 18 since the collected data covers an elevation range whereas Mabaage δ Odrip values are located at the 18 up to 1600 m (Chabrouh station) close to the basin center. δ Odrip values of Qadisha cave correspond to altitude of the studied caves and includes snowfall the highest part on both MWLs. Both Qadisha and 18 isotopic signals. This altitudinal trendline show a Jeita δ Odrip results of 2014 fall generally close to the 18 18 linear δ O-altitude relation of -0.13‰/100 m in West Lebanese MWL trend. Jeita δ Odrip results of 2011 Mount-Lebanon (Fig. 4B), i.e., a decrease in rainwater show a distinct displacement to the right of the Mount δ18O of 0.13‰ per 100 meters altitudinal increase. Lebanon MWL, that clearly indicates evaporation

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 68 Nehme et al. processes (Saad & Kazpard, 2007). The positive temperatures, which is the case for Jeita Cave (20°C), 18 δ Odrip values of the 2011 campaign in Jeita Cave (Fig. and less so for Qadisha Cave (9°C). Another possibility 4C) suggest several processes and factors that might is related to other cave environment parameters explain this particularity. Mickler et al. (2004) and Day which include enhanced ventilation (Muhlinghaus et and Handerson (2011) showed that the evaporation al., 2009; Deininger et al., 2012) that would lead to effect often occurs in caves with higher cave-air enhanced out-of-equilibrium processes.

Fig. 4. Summary of major cave drip and stream water δ18O and δ2H values plotted on the available meteoric water lines (MWL) with the location of their meteorological stations. A) location of the studied caves (yellow stars) and the meteorological stations (blue circles) in Saad et al. 2005 and Saad & Kazpard, 2007 and (orange circles) in Koeniger & Margane, 2014; B) the altitudinal trendline used in this study and derived from the Local MWL of 222 the Kelb basin (Koeniger & Margane, 2014); C) δ18O and δ2H values of 47 dripwater and stream samples plotted on the available MWLs. The δ18O values for stream waters in the Qadisha zones. Indeed, the Jeita underground stream exhibits and Jeita caves plot along the Mount Lebanon and a δ18O signal which is very close to the average δ18O Lebanese MWLs. The Qadisha stream water displays signal of the highest karstic springs feeding the δ18O values close to the drip water isotopic signal of Kelb basin: Nabaa el-Labane spring at 1647 m (avg. the same cave suggesting a similar water infiltration -7.26‰) and Nabaa al-Assal spring at 1528 m (avg. source for the vadose and the karst aquifer (phreatic) -7.32‰) (Aouad-Rizk et al., 2005; Koeniger et al., zones. However, the isotopic signal of Jeita stream 2017). The isotopic signals are also in agreement with exhibits higher values than the drip water isotopic the well-known infiltration basin (or recharge area) signal in Jeita upper cave advocating for different for the Jeita underground stream situated at a mean infiltration reservoir for the unsaturated and saturated altitude of 1669 m asl (Table 1).

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 Caves dripwater isotopic signals at Mount-Lebanon: implication for speleothem studies 69 The spread of the isotopic values cave waters along values are plotted first against altitude of the cave the Mount Lebanon MWL is related to the variability entrances (Fig. 5A). 18 2 18 18 in δ O and δ H in rainwater and therefore in cave drip Only the δ Odrip and δ Ostream values that fall closely water. This is mainly due to the spatial variability, on the local MWL were retained for the altitudinal inter-seasonal, or interannual variations in isotopic trend analysis (Fig. 5A). Figure 5A clearly shows a composition of rain. It suggests that all sampled poor altitudinal trend when considering only the dripwaters in the caves, especially in Jeita and less altitude of the cave entrance. However, there is a in Qadisha may be related to rainwater from different clear altitudinal trend with a regression coefficient seasons or even years depending on the residence R2 = 0.86 (P < 0.001) when considering the mean time of the water in the vadose, here epikarst zone. altitude of the infiltration basin (or recharge area) of The deuterium excess (d-excess) value is calculated these caves (Fig. 5B and 5C). Indeed, the basin from from δ18O values and δ2H using this equation: where water infiltrates is at higher elevation relative to the cave entrance due to the thick limestone overburden d-excess = δ2H – 8 * δ18O and the topography above the caves entrance. The 18 The d-excess, an indicator for the source and δ Owater values of Jeita stream are a clear example trajectories of atmospheric moisture (Rozanski, 1993; of that particularity, showing that the underground Sharp, 2007), is associated with evaporation at the water originates from an infiltration basin at a higher moisture source. The comparison of d-excessdrip vs altitude (avg. alt. 1669 m) than the cave entrance d-excessrain is necessary to understand if the measured (60 m) (Doummar, 2012; Koeniger et al., 2017). cave waters are controlled by a similar vapor source With the infiltration basin altitude (Table 1) taken than the rainwater. The Eastern Mediterranean source into consideration here, all points fall within the waters have d-excess values ranging from 14‰ to 95% confidence interval (Fig. 5B and 5C) except 19‰ (Kattan, 1997), whilst they reach 15‰ (Frot et one point, which represent cave waters taken from al., 2007) for Western Mediterranean sourced waters, Mabaage Cave at the end of August 2014. All drip and are close to 10‰ for Atlantic-sourced moisture. and stream waters δ18O values reach 0.2‰ per The Lebanese MWL exhibit a d-excess of 15.98‰ 100 m (Fig. 5B) with an overall offset of 0.07‰ which is within the range of the Eastern Mediterranean compared to the rainwater trend of 0.13‰ per 100 m waters’ values (Kattan, 1997). The d-excess for the calculated form the trendline (Fig. 4B) after Koeniger dripwater in cave indicate a value of 16.25, which is & Margane (2014). close to the one defined by the Lebanese MWL (Aouad- Comparable studies in the Mediterranean region Rizk et al., 2005; Saad & Kazpard, 2007; Koeniger & showed different small offsets between altitudinal Margane, 2014). gradients for precipitation and dripwater δ18O values. In the steep northern Italian Alps, eight caves Altitudinal trends in cave water δ18O aligned along two transects, show slightly different The Mediterranean air masses arriving from the gradients of 0.15 and 0.08‰ in dripwater (Johnston west are orographically uplifted as they reach the et al., 2013). In the eastern Adriatic coast and Dinaric Mount-Lebanon range. As the air rises and cools, mountains (), the offset reaches up to 0.2‰ the rainwater with a heavier isotope falls first, (Suric et al., 2016) similar to the offset measured in resulting in rainwater exhibiting more negative Mount-Lebanon cave waters. isotopic values with altitude (Bowen & Wilkinson, Clearly, the Δδ18O/100 m value of the dripwater 18 2002). Globally, the average change in δ Orain is (Fig. 5C) is site-specific, mainly due to: i) local 18 -0.2‰ per 100 m elevation gain (Rozanski et processes that influence the δ Odrip on its way to al., 1993). Locally, we determined this trend as the cave, such as those within the litter, soil, and -0.13‰/100 m (Fig. 4B). In order to understand epikarst (Beddows et al., 2016), ii) the relation to the altitude effect on the isotopic composition of the precipitation altitude gradient (rainfall quantity, cave water (drip and stream), the δ18O cave water patterns, and frequencies above the cave infiltration

18 18 Fig. 5. Altitudinal trends in cave water δ O (drips and streams). A) Adjusted linear regression between the δ Ocave water and the altitude of the cave 18 18 entrance; B) δ Ocave water and the altitude of the infiltration basin; C) Plot showing the δ Odripwater trendline only vs the infiltration basin altitude. The calculated interval of confidence (dashed line) for each linear regression is 95% and the significance p-value for all three graphs is P < 0.001.

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 70 Nehme et al. basin), or iii) the hilly topography above the cave and cave, while in summer seasons, 18O (2H)-enriched the limits of the calculated infiltration basin defined water will partially evaporate in the unsaturated zone, herein. especially when shallow overburden exists above the Figure 6A compares the offset between altitudinal cave (Wackerbarth et al., 2010, 2012). The cave waters gradients for precipitation and dripwater δ18O values. are therefore normally biased towards lower/lighter Within the limit of the 95% confidence interval of δ18O/δ2H values compared to the rainwater isotopic the precipitation δ18O trendline and considering the signal (Wackerbarth et al., 2012). 2σ sigma error of the dripwater, the drip δ18O values Generally, cave dripwaters in Lebanon are mostly fall generally close to the precipitation δ18O trendline the result of percolation happening during the wet except for Qadisha Cave. There is, however, a minor season (from autumn to spring snowmelt), with a negative offset (0.2‰ at low altitude to 1.2‰ at high longer infiltration period at higher altitudes due to altitude) between the dripwater and the precipitation snowmelt. For Qadisha Cave, which is located at a δ18O trendline. higher altitude, the offset between the precipitation This offset, similar to the one observed for the δ18O trendline and the dripwaters isotopic signals Adige, Valsugana valleys, northern Italy (Johnston is the most negative when compared to the other et al., 2013), the Adriatic coast (Suric et al., 2016), studied caves. This is explained by the infiltration of and Vancouver, Canada (Beddows et al., 2016) winter water enhanced by a negative isotopic value represents a bias due to the infiltration effects of winter snow, especially at higher altitudes (Aouad- of rainwater. In winter, the infiltrated water from Rizk et al., 2005) and contributing into the vadose rainfall/snowmelt with lower δ18O values reaches the water budget.

18 18 18 Fig. 6. Cave dripwaters δ O values in Lebanese caves compared to δ Orain: A) trendline showing the mean δ Odripwater (blue dots) of each cave (this study) compared to the altitudinal trendline used in this study (red rectangles) and derived from the Local MWL (δ18O weighted-mean rainwater values) after Koeniger & Margane (2014). The calculated interval of confidence (dashed line) for the δ18O weighted-mean rainwater 18 regression is 95% and the significance p-value is P < 0.002; B) δ Odripwater measured at nine different sampling sites (JC-01to 09) in Jeita Cave on a yearly basis (data in Koeniger & Margane, 2014). Regarding the altitudinal effect on the 18O- and 60–90% compared to the Victoria rainfall records of 2 18 H-depleted dripwater, our study shows that δ Odrip the same year. In Villars, Chauvet, and Orgnac caves, 18 values decrease up to 3‰ between Kanaan and southern France, the δ Odrip values stayed stable for Qadisha caves (Fig. 6A). This amplitude attributed to 15 years with little seasonal variations compared 18 the altitudinal effect could theoretically be increased to drip rate measurement. In Lebanon, our δ Odrip 18 by variations in the δ Odrip values related to site- data, even though stable during the autumn season specific characteristics or to a seasonal bias between prevent us from assessing a seasonal variability for winter and summer dripwaters values transferred by all four cave sites. However, a previous campaign on the rainwater seasonal variations. Indeed, rainwater dripwater isotopic measurement completed at nine δ18O values in Lebanon show clearly a seasonal bias drip sites in Jeita Cave (Koeniger & Margane, 2014) with a variation up to 5‰ (Saad & Kazpard, 2007; show little variability at each drip site over a complete Koeniger & Margane, 2014) between early-winter and rainy and early-summer season (Fig. 6B), but rather a winter-spring seasons. In fact, seasonal variations spatial variability between each drip site. Indeed, the in meteoric precipitation may range up to >15‰ maximum seasonal variability of 1‰ is only recorded 18 18 for δ O (Genty et al., 2014). However, the majority in JC-05 site (Fig. 6B). The δ Odrip yearly average of cave sites studied around the world (Genty et al., is -5.24‰ in Jeita cave showing a low seasonal 2014; Beddows et al., 2016) demonstrated that drip variability with a standard deviation of ±0.48. waters typically show little or no isotopic seasonality Therefore, the seasonal variations in cave dripwaters 18 compared to the variations in meteoric precipitation. as seen in Jeita δ Odrip measurement, account less in 18 For instance, the δ Odrip variability of caves in the altitudinal effect on the lowering of the dripwater Vancouver Island, Canada is reduced in amplitude by isotopic values.

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 Caves dripwater isotopic signals at Mount-Lebanon: implication for speleothem studies 71 Implications for future speleothems-based and calcite monitoring with automated logging paleoclimate studies equipment (pCO2, temperature, humidity, etc.) The isotopic signal of the dripwater in Lebanese will continue to refine the interpretations that caves located on the western flank of Mount-Lebanon have been based on the initial monitoring findings falls generally on the local MWL (Koeniger & Margane, presented here. 2014) as well as the Mount-Lebanon MWL (Aouad- • To build further on this study, the altitudinal Rizk et al., 2005). This implies: i) an identical source trend signal should be confirmed by modern of water being derived from rain forming over the calcite from the same caves, in which the trend Mediterranean basin as indicated by similar d-excess should have a similar gradient. values of the water, ii) a reduced evapotranspiration effect, observable on only some samples with a ACKNOWLEDGMENTS clear offset to the right of the MWL, probably due to increased cave ventilation, and iii) a longer infiltration This study was funded by the 2014 mobility period occurring in the unsaturated zone at higher fellowship program of the Belgian Federal Scientific altitudes. Policy (BELSPO), co-funded by the Marie Curie Whilst some exceptions might occur as seen Actions of the European Commission, and the in some drip water in Jeita Cave during the 2011 European Union’s Horizon 2020 Research. Laboratory campaign, which were more exposed to ventilation at analysis conducted at LSCE-Paris were funded under some locations, most of the cave dripwater exhibits the MISTRALS research program. We acknowledge a similar δ18O and δ2H signal as the local rainwater. the assistance of St-Joseph University of Beirut for However, a slight offset towards lower ‘winter values’ facilitating the access to caves with the help of ALES may occur due to a preferential water recharge during (Association Libanaise d’Etudes Spéléologiques) and winter months, including the recharge by melting SCL (Spéléo-Club du Liban) and the support of both snow. club members who collected water and calcite Regarding the altitudinal trend observed in the samples during field campaigns. The authors are rainwater over the Mount-Lebanon range (Fig. 4B), grateful to Bogdan Onac for his editorial handling of the isotopic signal in dripwater exhibits an altitudinal the manuscript and endless patience and especially trend, but with a slightly different gradient (-0.21‰ to the two anonymous reviewers for their constructive per 100 m) than the rainwater (-0.13‰ per 100 m). comments and revisions that considerably improved This is however, more significant when the dripwater the quality of the manuscript. isotopic signal is compared to the altitude of the infiltration basin of each cave (Fig. 5C). REFERENCES

CONCLUSION Abi-Saleh B. & Safi S., 1988 – Carte de la végétation du Liban. Ecologia Mediterranea, 14: 123-141. The preliminary dripwater isotopic measurements Abou Zakhem A. & Hafez R., 2010 – Climatic factors and temperature conducted on four caves located in controlling chemical and isotopic characteristics of the western flank of Mount-Lebanon, revealed the precipitation in Syria. Hydrological Processes, 24: following important conclusions for future speleothem- 2641-2654. https://doi.org/10.1002/hyp.7646 based interpretation of paleoclimate changes at both Aouad-Rizk A., Job J.O., Khalil S., Touma T., Bitar C., local and regional scales: Bocquillon C. & Najem W., 2005 – δ18O and δ 2H contents • Despite for some water samples influenced by over Mount-Lebanon related to mass trajectories and evaporative processes, the drip water exhibits local parameters. In: Isotopic composition of precipitation isotopic values in agreement with the local in the Mediterranean Basin in relation to air circulation rainwater. Therefore, for paleoclimatic patterns and climate. IAEA-TECDOC, 1453: 75-82. reconstructions (or fluid inclusion analysis) Alpert P., Price C., Krichak S.O., Ziv B., Saaroni H., should be preferentially chosen outside a possible Osetinsky I. & Kishcha P., 2005 – Tropical teleconnections ventilation-influenced area of the cave. to the Mediterranean climate and weather. Advances in • The altitudinal trend confirmed previously in the Geosciences, 2: 157-160. rainwater isotopic composition on the western https://doi.org/10.5194/adgeo-2-157-2005 flank of Mount-Lebanon is demonstrated also in Baldini J.U.L., McDermott F. & Fairchild I.J., 2006 – cave drip water indicating the transfer to the cave Spatial variability in cave drip water hydrochemistry: through the vadose zone of the spatial isotopic Implications for stalagmite paleoclimate records. signals of the rainwater. The isotopic composition Chemical Geology, 235: 390-404. https://doi.org/10.1016/j.chemgeo.2006.08.005 of the dripwaters, however, exhibits a slightly Baker A., Barnes W. & Smart P., 1997 – Variations in the higher negative δ18O/100 m gradient for cave drip discharge and organic matter content of stalagmite drip water due to slower infiltration of winter waters. waters in Lower Cave, Bristol. Hydrological Processes, The isotopic dripwater signal represents therefore 11: 1541-1555. mostly a lower limit of the isotopic signal of the https://doi.org/10.1002/(SICI)1099-1085(199709)11: corresponding rain/snow melt. 11%3C1541::AID-HYP484%3E3.0.CO;2-Z • The results of this study can further help in the Bar-Matthews M., Ayalon A., Matthews A., Sass E. & Halicz interpretation of past altitudinal trends based L., 1996 – Carbon and oxygen isotope study of the active on speleothems. Additional future cave water water-carbonate system in a karstic Mediterranean

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 72 Nehme et al. cave: implications for paleoclimate research in semiarid Edgell H.S., 1997 – Karst and hydrogeology of Lebanon. regions. Geochimica Cosmochimica Acta, 60: 337-347. Carbonates and Evaporites, 12 (2): 220-235. https://doi.org/10.1016/0016-7037(95)00395-9 https://doi.org/10.1007/BF03175419 Bar-Matthews M., Ayalon A., Gilmour M., Matthews Frot E., Van Wesemael B., Vandenschrick G., Souchez A. & Hawkesworth C.J., 2003 – Sea-land oxygen R. & Benet A.S., 2007 – Origin and type of rainfall isotopic relationships from planktonic foraminifera and for recharge of a karstic aquifer in the western speleothems in the Eastern Mediterranean region and Mediterranean: a case study from the Sierra de Gador- their implication for paleorainfall during interglacial Campo de Dalias (southeast Spain). Hydrological intervals. Geochimica Cosmochimica Acta, 67: 3181-3199. Processes, 21: 359-368. https://doi.org/10.1016/S0016-7037(02)01031-1 https://doi.org/10.1002/hyp.6238 Beddows P.A., Mandic M., Ford D.C. & Schwarcz H.P., Fairchild I.J. & Baker A., 2012 – Speleothem science: 2016 – Oxygen and hydrogen isotopic variations from process to past environments. Wiley-Blackwell, between adjacent drips in three caves at increasing Chichester, 450 p. elevation in a temperate coastal rainforest, Vancouver https://doi.org/10.1002/9781444361094 Island, Canada. Geochimica Cosmochimica Acta, 172: Friedman I. & O’Neil J.R., 1977 – Data of geochemistry: 370-386. https://doi.org/10.1016/j.gca.2015.08.017 Compilation of stable isotope fractionation factors of Bowen G.J. & Wilkinson B., 2002 – Spatial distribution geochemical interest. United States Geological Survey, 18 of δ O in meteoric precipitation. Geology, 30: 315-318. Washington, USA, 170 p. https://doi.org/10.1130/0091-7613(2002)030%3C03 Fohlmeister J., Scholz D., Kromer B. & Mangini A., 2011 15:SDOOIM%3E2.0.CO;2 – Modelling carbon isotopes of carbonates in cave drip Burlet C., Vanbrabant Y., Piessens K., Welkenhuysen K. water. Geochimica Cosmochimica Acta, 75: 5219-5228. & Verheyden S., 2015 – Niphargus: a silicon band-gap https://doi.org/10.1016/j.gca.2011.06.023 sensor temperature logger. Computers & Geosciences, Ford D. & Williams P., 2007 – Karst hydrogeology and 74: 50-59. geomorphology. John Wiley & Sons, Chichester, 562 p. https://doi.org/10.1016/j.cageo.2014.10.009 https://doi.org/10.1002/9781118684986 Cheng H., Sinhac A., Verheyden S., Nader F.H., Li X.L., Frisia S. & Borsato A., 2010 – Karst. In: Alonso-Zarza A.M. Zhang P.Z., Yin J.J., Yi L., Peng Y.B., Rao Z.G., Ning & Tanner L.H. (Eds.), Developments in sedimentology, Y.F. & Edwards R.L., 2015 – The climate variability in carbonates in continental settings. Elsevier, Amsterdam, northern Levant over the past 20,000 years. Geophysical p. 269-318. Research Letters, 42: 8641-8650. https://doi.org/10.1016/S0070-4571(09)06106-8 https://doi.org/10.1002/2015GL065397 Gat J., 1980 – The isotopes of hydrogen and oxygen Day C.C. & Henderson G.M., 2011 – Oxygen isotopes in precipitation. In: Fritz P. & Fontes J.Ch. (Eds.), in calcite grown under cave-analogue conditions. Handbook of environmental isotope geochemistry, 1: Geochimica Cosmochimica Acta, 75: 3956-3972. 22-48. https://doi.org/10.1016/j.gca.2011.04.026 Gat J.R., Klein B., Kushnir Y., Roether W., Wernli H., Dansgaard W., 1964 – Stable isotopes in precipitation. Yam R. & Shemesh A., 2003 – Isotope composition of Tellus, 16: 438-468. air moisture over the Mediterranean Sea: an index of the https://doi.org/10.3402/tellusa.v16i4.8993 air–sea interaction pattern. Tellus, Series B-Chemical & Deininger M., Fohlmeister J., Scholz D. & Mangini A., Physical Meteorology, 55 (5): 953-965. 2012 – Isotope disequilibrium effects: The influence of https://doi.org/10.1034/j.1600-0889.2003.00081.x evaporation and ventilation effects on the carbon and Gat J.R., Ben-Mair R., Yama R., Yakir D. & Wernli H., oxygen isotope composition of speleothems - A model 2005 – The Isotope Composition of Atmospheric Waters approach. Geochimica Cosmochimica Acta, 96: 57-79. https://doi.org/10.1016/j.gca.2012.08.013 in Israel’s Coastal Plain. In: Isotopic composition of De Bondt K., Seveno F., Petrucci G., Rodriguez F., precipitation in the Mediterranean Basin in relation to Joannis C. & Claeys P., 2018 – Potential and limits of air circulation patterns and climate. IAEA-TECDOC, stable isotopes (δ18O and δ2H) to detect parasitic water in 1453: 99-114. sewers of oceanic climate cities. Journal of Hydrology: Genty D., Blamart D., Ghaleb B., Plagnes V., Causse Regional Studies, 18: 119-142. C.h., Bakalowicz M., Zouari K., Chkir N., Hellstrom J., https://doi.org/10.1016/j.ejrh.2018.06.001 Wainer K. & Bourges F., 2006 – Timing and dynamics Dirican A., Unal S., Acar Y. & Demircan M., 2005 – of the last deglaciation from European and North African 13 The temporal and seasonal variation of δ 2H and δ 18O δ C stalagmite profiles– comparison with Chinese and in atmospheric water vapour and precipitation from South Hemisphere stalagmites. Quaternary Science Ankara, Turkey in relation to air mass trajectories at Reviews, 25: 2118-2142. Med. Basin. In: Isotopic composition of precipitation in https://doi.org/10.1016/j.quascirev.2006.01.030 the Mediterranean Basin in relation to air circulation Genty D., Labuhn I., Hoffmann G., Danis P.A., Mestre patterns and climate. IAEA-TECDOC, 1453: 191-214. O., Bourges F., Wainer K., Massault M., Van Exter S., Doummar J., 2012 – Identification of indicator parameters Régnier E., Orengo Ph., Falourd S. & Minster B., 2014 for the quantitative assessment of vulnerability in karst – Rainfall and cave water isotopic relationships in two aquifers. Unpublished Thesis, University of Gottingen, South-France sites. Geochimica Cosmochimica Acta Gottingen, 116 p. 131: 323-343. Dubertret L., 1975 – Introduction à la carte géologique au https://doi.org/10.1016/j.gca.2014.01.043 1/50000e du Liban. Notes et Mémoires sur le Moyen- Hellstrom J., McCulloch M. & Ston J., 1998 – A detailed Orient, 23: 345-403. 31.000-year record of climate and vegetation change, El-Asrag A.M., 2005 – Effect of synoptic and climatic from the isotope geochemistry of two New Zealand situations on fractionation of stable isotopes in rainwater speleothems. Quaternary Research, 50: 167-178. over Egypt and east Mediterranean. In: Isotopic https://doi.org/10.1006/qres.1998.1991 composition of precipitation in the Mediterranean Basin Hendy C.H., 1971 – The isotopic geochemistry of in relation to air circulation patterns and climate. IAEA- speleothems-I: The calculations of the effects of TECDOC, 1453: 51-73. different modes of formation on the isotopic composition

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 Caves dripwater isotopic signals at Mount-Lebanon: implication for speleothem studies 73 of speleothems and their applicability as paleoclimate in southeastern Australia. Hydrological Processes, 21: indicators. Geochimica Cosmochimica Acta, 35: 801-824. 1737-1748. https://doi.org/10.1002/hyp.6356 https://doi.org/10.1016/0016-7037(71)90127-X Mickler P.J., Banner J.L., Stern L., Asmerom Y., Edwards Hendy C.H., Wilson A.T., Popplewell K.B. & House R.L. & Ito E., 2004 – Stable isotope variations in modern D.A., 1977 – Dating of geochemical events in Lake tropical speleothems: Evaluating equilibrium vs. kinetic Bonney, Antarctica, and their relation to glacial and isotope effects. Geochimica Cosmochimica Acta, 68: climate changes. New Zealand Journal of Geology & 4381-4393. Geophysics, 20 (6): 1103-1122. https://doi.org/10.1016/j.gca.2004.02.012 https://doi.org/10.1080/00288306.1977.10420698 Miorandi R., Borsato A., Frisia S., Fairchild I.J. & Richter Hakim B., 1985 – Recherche hydrologiques et hydrochimiques D.K., 2010 – Epikarst hydrology and implications for sur quelques méditerranéens: Liban, Syrie, Maroc. stalagmite capture of climate changes at Grotta di PhD thesis, Lebanese University Beirut, Lebanon, 700 p. Ernesto (NE Italy): results from long-term monitoring. Hakim B. & Karakabi S., 1988 – Coloration du gouffre Hydrological Processes, 24: 3101-3114. de Faouar Dara et de la grotte de Kassarat. Al-Ouate- https://doi.org/10.1002/hyp.7744 Ouate Spéléo-Club du Liban, 3: 18-33. Mhawej M., Faour G., Fayad A. & Shaban A., 2014 – International Atomic Energy Agency/World Towards an enhanced method to map snow cover areas Meteorological Organization (2008) Global network of and derive snow-water equivalent in Lebanon. Journal isotopes in precipitation. The GNIP Database. of Hydrology, 513: 274-282. https://nucleus.iaea.org/Pages/GNIPR.aspx https://doi.org/10.1016/j.jhydrol.2014.03.058 (Accessed date: May 2016) Muhlinghaus C., Scholz D. & Mangini A., 2009 – Jabbour-Gedeon B. & Zaatar J., 2013 – Mgharet Modelling fractionation of stable isotopes in stalagmites. Mabaage: nouvelles découvertes et aménagement Geochimica Cosmochimica Acta, 73: 7275-7289. touristique “surprenant”. Spéléorient, Association https://doi.org/10.1016/j.gca.2009.09.010 libanaise d’Etudes spéléologiques, 6: 61-67. Nehme C., Jabbour-Gedeon B., Gerard P.-C., Sadier B. Johnston V.E., Borsato A., Spötl C., Frisia S. & Miorandi & Delannoy J.J., 2009 – Reconstitution sepéléogénique R., 2013 – Stable isotopes in caves over altitudinal de la grotte de Kanaan (Antelias, Liban): contribution à la gradients: fractionation behaviour and inferences for morphogenèse du Nahr Antélias. Karstologia, 54: 21-30. speleothem sensitivity to climate change. Climate of the Nehme C., Verheyden S., Noble S.R., Farrant A.R., Sahy Past, 9: 99-118. D., Hellstrom J., Delannoy J.J. & Claeys P., 2015 – https://doi.org/10.5194/cp-9-99-2013 Reconstruction of MIS 5 climate in the central Levant Karam F., 2002 – Climate change and variability in using a stalagmite from Kanaan Cave, Lebanon. Climate Lebanon: impact on land use and sustainable agriculture of the Past, 11: 1785-1799. development. Proceedings of the 1st technical workshop https://doi.org/10.5194/cp-11-1785-2015 of the “Mediterranean” component of CLIMAGRI project Nehme C., Verheyden S., Breitenbach S.F., Gillikin D.P., on climate change and agriculture, FAO, Rome, 25-27. Verheyden A., Cheng H., Edwards R.L., Hellstrom J., Karkabi S., 1990 – Cinquantenaire de la spéléologie Noble S.R., Farrant A.R. & Sahy D., 2018 – Climate libanaise. Al-Ouate-Ouate, special edition, 138 p. dynamics during the penultimate glacial period recorded Kattan Z., 1997 – Environmental isotope study of the in a speleothem from Kanaan Cave, Lebanon (central major karst springs in Damascus limestone aquifer Levant). Quaternary Research, 90 (1): 1-16. systems: case of the Figeh and Barada springs. Journal https://doi.org/10.1017/qua.2018.18 of Hydrology, 193 (1-4): 161-182. Peel M.C., Finlayson B. & McMahon Th., 2007 – Updated https://doi.org/10.1016/S0022-1694(96)03137-X world map of the Köppen-Geiger climate classification. Koeniger P. & Margane A., 2014 – Stable isotope Hydrological and Earth Systems Sciences Discussions, investigations in the Jeita Spring catchment. BGR, 4 (2): 439-473. Special report on protection of Jeita Spring, 12, 48 p. https://doi.org/10.5194/hessd-4-439-2007 Koeniger P., Margane A., Abi-Rizk J. & Himmelsbach T., Rozanski K., Araguas-Araguas L. & Gonfiantini R., 2017 – Stable isotope based mean catchment altitudes 1993 – Isotopic patterns in modern global precipitation. of springs in the Lebanon Mountains. Hydrological Geophysical Monographical Series, 78: 1-36. Processes, 21: 3708-3718. https://doi.org/10.1029/GM078p0001 https://doi.org/10.1002/hyp.11291 Saad Z., V. Kazpard A.G. Samrani El. & Slim K., 2005 Luetscher M. & Jeannin P.-Y., 2004 – Temperature – Chemical and isotopic composition of rainwater in distribution in karst systems: the role of air and water coastal and highland regions in Lebanon. Journal of fluxes. Terra Nova, 16: 344-350. Environmental Hydrology, 13: 1-11. https://doi.org/10.1111/j.1365-3121.2004.00572.x Saad Z. & Kazpard V., 2007 – Seasonal effect on the Lachniet M.S., 2009 – Climatic and environmental controls isotopic pattern of rainwater in Lebanon. Journal of on speleothem oxygen isotope values. Quaternary Environmental Hydrology, 15: 1-15. Science Reviews, 28: 412-432. Sharp Z., 2007 – Principles of stable isotope geochemistry. https://doi.org/10.1016/j.quascirev.2008.10.021 Pearson Prentice Hall, New Jersey, 344 p. Mattey D., Fairchild I., J. Atkinson T.C., Latin J.-P., Scholz D., Muhlinghaus C. & Mangini A., 2009 – Ainsworth M. & Durell R., 2010 – Seasonal microclimate Modelling δ13C and δ18O in the solution layer on control of calcite fabrics, stable isotopes and trace stalagmite surfaces. Geochimica Cosmochimica Acta, elements in modern speleothem from St Michaels 73: 2592-2602. Cave, Gibraltar. In: Pedley H.M. & Rogerson M. (Eds.), https://doi.org/10.1016/j.gca.2009.02.015 and speleothems: Unravelling the microbial and Spotl C., Fairchild I.J. & Tooth A.F., 2005 – Cave air physical controls. Geological Society, London, Special control on dripwater geochemistry, Obir Caves (Austria): Publications, 336: 323-344. Implications for speleothem in dynamically https://doi.org/10.1144/SP336.17 ventilated caves. Geochimica Cosmochimica Acta, 69: McDonald J. & Drysdale R., 2007 – Hydrology of cave drip 2451-2468. waters at varying bedrock depths from a karst system https://doi.org/10.1016/j.gca.2004.12.009

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019 74 Nehme et al. Suric M., Loncaric R., Loncar N., Buzjak N., Bajo P. & Verheyden S., Nader F.H., Cheng H.J., Edwards L.R. & Drysdale R., 2017 – Isotopic characterization of cave Swennen R., 2008b – Paleoclimate reconstruction in environments at varying altitudes on the eastern Adriatic the Levant region from the geochemistry of a coast (Croatia) – Implications for future speleothem- stalagmite from the Jeita cave, Lebanon. Quaternary based studies. Journal of Hydrology, 545: 367-380. Research, 70: 368-381. https://doi.org/10.1016/j.jhydrol.2016.12.051 https://doi.org/10.1016/j.yqres.2008.05.004 Tawk J. & Kaasamani H., 2014 – La grotte de Qadisha Wackerbarth A., Langebroek P.M., Werner M., Lohmann - première grotte touristique du Liban. Al-Ouate-Ouate, G., Riechelmann S., Borsato A. & Mangini A., 2012 Spéléo-club du Liban, 14: 36-41. – Simulated oxygen isotopes in cave drip water and Tremaine D.M., Froelich P.N. & Wang Y., 2011 – Speleothem speleothem calcite in European caves. Climate of the calcite farmed in situ: Modern calibration of δ18O and Past, 8: 1781-1799. δ13C paleoclimate proxies in a continuously monitored https://doi.org/10.5194/cp-8-1781-2012 Wackerbarth A., Scholz D., Fohlmeister J. & Mangini natural cave system. Geochimica Cosmochimica Acta, A., 2010 – Modelling the δ18O value of cave drip water 75: 4929-4950. and speleothem calcite. Earth and Planetary Science https://doi.org/10.1016/j.gca.2011.06.005 Letters, 299 (3-4): 387-397. Van Geldern R. & Barth-Johannes A.C., 2012 – https://doi.org/10.1016/j.epsl.2010.09.019 Optimization of instrument setup and post-run corrections White B., 1988. Geomorphology and hydrology of karst for oxygen and hydrogen stable isotope measurements terrains. Oxford University Press, New York, 464 p. of water by isotope ratio infrared spectroscopy (IRIS). Ziv B., Dayan U., Kuschnir Y., Roth C., Enzel Y., 2006 Limnology Oceanography Methods, 10: 1024-1036. – Regional and global atmospheric patterns governing https://doi.org/10.4319/lom.2012.10.1024 rainfall in the southern Levant. International Journal of Verheyden S., Genty D., Deflandre G., Quinif Y. and Climatology, 26: 55-73. Keppens E., 2008a – Monitoring climatological, https://doi.org/10.1002/joc.1238 hydrological and geochemical parameters in the Père Ziv B., Saaroni H., Romem M., Heifetz E., Harnik N. Noël cave (Belgium): implication for the interpretation & Baharad. A., 2010 – Analysis of conveyor belts in of speleothem isotopic and geochemical time-series. winter Mediterranean cyclones. Theoretical Applied International Journal of Speleology, 37 (3): 221-234. Climatology, 99: 441-455. https://doi.org/10.5038/1827-806X.37.3.6 https://doi.org/10.1007/s00704-009-0150-9

International Journal of Speleology, 48 (1), 63-74. Tampa, FL (USA) January 2019